It is difficult to eliminate profanity, racist comments and other offensive and anti-social behavior within chat using automated means such as word filters because there are many variants to profanity and ways to foil contextually insensitive filtering (inserting symbols or extraneous letters into profanity, coming up with new terminology to refer to targeted groups, etc.)
The aforementioned behavior within chat can and does occur within chat environments tied to online gaming. In this context, attempts to deter this type of behavior typically requires manual administrator intervention and/or tools as mentioned above which can be largely ineffective.